Research on EEG Classification with Neural Networks Based on the Levenberg-Marquardt Algorithm

In the brain-computer interface (BCI), the feature extraction and classification of electroencephalogram (EEG) can be achieved by massive study of the Multilayer feedforward neural network. But the BP neural network based on error back propagation converges slowly, and has low efficiency in training, limited accuracy in classification. To solve these problems, the quick and stable Levenberg-Marquardt algorithm is adopted instead of the BP algorithm to train the neural network. The MATLAB simulation experiment of classifying the EEG signals of the motor imagery of left hand and right hand uses the Graz data set B from the BCI competition 2008. The simulation results show that the accuracy rate of this algorithm is 87.1%, which is superior to 78.2% of the BP algorithm, and it converges better as well. This technology provides an effective way to EEG classification.